Estrogen Receptor-␤ Expression in Human Testicular Germ Cell Tumors

Home , ER (season 10), ER (season 11), ER (season 13), ER (season 14), ER (season 15), Testicular cancer

Vol. 9, 4475–4482, October 1, 2003 Clinical Cancer Research 4475

Vernon Pais,1 Irwin Leav, Kin-Mang Lau,2 gene, was localized to spermatagonia of the normal testis, Zhong Jiang, and Shuk-Mei Ho3 but its expression dramatically reduced in seminomas. With the exception of spermatagonia, androgen receptor was Division of Urology, Department of Surgery [V. P., K-M. L., S-M. H.], and Departments of Pathology [I. L., Z. J.], Cell Biology found in all of the germ cells of the normal testis, but, aside [S-M. H.], and Physiology [S-M. H.], University of Massachusetts from trace staining in 3 of 5 endodermal sinus tumor cells, Medical School, Worcester, Massachusetts 01605, and Department of it was not detected immunohistochemically in any other Pathology, Tufts University School of Veterinary Medicine, Grafton, germ cell cancer. Massachusetts [I. L.] Conclusions: We confirm expression of ER-␤, but not ER-␣, in normal testicular cells, suggesting that only the ABSTRACT former ER subtype mediates the action of estrogen in the Purpose: Estrogen exposure has been linked to a risk human male gonad. Our results provide the first evidence ␤ for the development of testicular germ cell cancers. The that only ER- is expressed in testicular germ cell tumors. effects of estrogen are now known to be mediated by estro- Its expression is down-regulated in seminomas and embry- gen receptor (ER)-␣ and -␤ receptor subtypes, but only onal cell carcinomas but remains high in endodermal sinus ␤ ER-␤ has been found in human normal testis. The goal of tumors and in teratomas. The observed differences in ER- the present study was to compare the localization and ex- expression levels among different testicular germ cell pression levels of these ER subtypes in testicular germ cell tumors may reflect divergent pathways of differentiation/ cancers (seminomas and nonseminomatous germ cell tu- dedifferentiation of these neoplasms from a common mors) with normal testis. For completeness, expression of precursor. Collectively, these findings provide a possible androgen and progesterone receptors was also investigated. mechanistic link between estrogen exposure and testicular Experimental Design: Immunohistochemistry was used cancer risk. to localize the expression of steroid receptors in 39 archival testicular germ cell cancers and 5 morphologically normal INTRODUCTION testes. Expression of the steroid receptors at the transcript Testicular germ cell tumors are the most common malig- level was semiquantified by reverse transcription-PCR in nancy in males between 15 and 34 years of age, and represent a 5 paired fresh-frozen specimens of normal and neoplastic major cause of death attributable to cancer in this age group (1). testes. Every year, ϳ7400 new cases of testicular tumors are diagnosed Results: ER-␣ was not expressed in the human normal in the United States. The incidence of this type of cancer has testis. It was also absent in all of the testicular germ cell increased progressively throughout the twentieth century (2). cancers studied. In contrast, ER-␤ was strongly expressed in Germ cell tumors can be subdivided into seminoma and various germ cells of the normal testis. However, its expres- NSGCTs,4 which consist of embryonal cell carcinoma, chorio- sion was markedly diminished in seminomas, embryonal cell carcinoma, yolk sac tumor, and teratoma. Neoplasms that con- carcinomas, and in mixed germ cell tumors, at both tran- tain more than one tumor cell components, e.g. seminoma and scriptional and translational levels. In contrast, ER-␤ re- embryonal cell carcinoma, are referred to as mixed germ cell mained highly expressed in endodermal sinus tumors and tumors. Seminoma and NSGCTs not only present with distinc- teratomas. Progesterone receptor, an estrogen-regulated tive clinical features, they also differ with respect to therapy and prognosis (3). Whereas the etiology of testicular germ cell cancers remains undefined, exposure to certain hormones (in particular estrogen) at the time of testicular differentiation in utero has Received 12/31/02; revised 5/16/03; accepted 5/28/03. The costs of publication of this article were defrayed in part by the long been implicated as a risk factor for developing these payment of page charges. This article must therefore be hereby marked neoplasms (4). An earlier case-control study (5) demonstrated advertisement in accordance with 18 U.S.C. Section 1734 solely to first trimester exposure of the mother to exogenous estrogen indicate this fact. produced a 8-fold rise in risk for testicular cancer in the son. This work was supported in part by NIH Grants CA15776 and DK610840, and a United States Army Prostate Cancer Program Grant More recently, other investigations (6, 7) confirmed increased DAMD17-98-1-8606 (to S-M. H.). 1 Present address: Department of Urology, St. Elizabeth’s Medical Cen- ter of Boston, Boston, MA 02135. 2 Present address: Department of Anatomical and Cellular Pathology, Prince of Wales Hospital, Shatin, New Territories, Hong Kong, China. 4 The abbreviations used are: NSGCT, nonseminomatous germ cell 3 To whom requests for reprints should be addressed, at Room S4-746, tumor; ER, estrogen receptor; PR, progesterone receptor; AR, androgen Department of Surgery, University of Massachusetts Medical School, 55 receptor; RT-PCR, reverse transcription-PCR; GAPDH, glyceralde- Lake Avenue North, Worcester, MA 01655. Phone: (508) 856-1909; hyde-3-phosphate dehydrogenase; IGCNU, intratubular germ cell neo- Fax: (508) 856-8699; E-mail: [email protected]. plasia of unclassified type.

incidences of testicular germ cell tumors in men with maternal logical studies, conducted to localize ER in testicular tumors, exposure to diethylstilbestrol. Furthermore, a higher incidence used only antibodies directed against the ␣ subtype (28). of testicular cancer was observed among men exposed prena- In the current investigation we compared the expression tally to excess maternal hormone as in the case of preterm birth, ER-␤, ER-␣, AR and PR in 39 human testicular germ cell being a twin, and first births among young mothers (6, 8). In malignancies with that found in normal testicular specimens. contrast, a protective effect was noted with in utero exposure to ER-␤ expression was studied at the immunohistochemical level presumed low-estrogen states, such as those associated with using a well-characterized, ER subtype-specific antibody, di- heavy cigarette use and bleeding/threatened miscarriage (6). rected against the F domain of the human receptor (29). Fur- Adult exposure to excess estrogen as in conditions associated thermore, expression of these receptors was studied at the tran- with development of gynecomastia (9), treatment of prostate script level in 5 testicular germ cell cancers and results cancer with estrogen (10), or occupational contact with estrogen compared with expression levels in morphologically normal mimics (e.g. organochlorines; Refs. 11, 12) have all been linked testicular tissues. to a higher risk for testicular cancers. Experimentally, prenatal and postnatal administration of estrogens consistently causes testicular tumor development in some strains of rodents (13). MATERIALS AND METHODS Despite the above-cited epidemiological reports linking Testicular Tissues. Thirty-nine archival cases from rad- estrogens to testicular cancer the mechanisms by which estro- ical orchiectomies and 5 from simple orchiectomies (ages 21– gens contribute to the pathogenesis of these neoplasms remains 57) were selected from the surgical pathology files of the University of Massachusetts Medical School. These specimens unclear. It has been reported that serum estrogen levels are were collected during a 6-month period (at the end of 2000 and elevated in patients with testicular germ-cell cancers (14, 15), a the beginning of 2001). The collection included 17 seminomas, condition which likely resulted from local production by the and 22 NSGCTs consisting of 10 mixed germ cell tumors with tumor tissue, because aromatase activity has been demonstrated embryonal cell components, 5 with endodermal sinus compo- in Leydig cells, Sertoli cells, and various germ cells of the nents, 3 cases of embryonal carcinoma, and 4 cases of mature normal testis (16, 17). A recent report suggests that the carci- teratomas. Serial sections, 5–6 ␮m, were cut from paraffin- nogenic effects of estrogen on testicular cells may involve embedded specimens, mounted on glass slides, and stored un- ER-mediated oxidative DNA damage (18). In this study, expo- baked until immunohistochemistry was performed. ␣ sure of rat testicular cells to 17 -ethinylestradiol caused in- In addition, transcript expression was studied with mRNA Ј creased formation of 7, 8-dihydro-8-oxo-2 -dexoyguanosine, a extracted from 5 simple orchiectomy surgical specimens and marker of oxidative DNA-damage, which could be blocked by compared those in extracts from morphologically “normal” tes- cotreatment with a pure ER antagonist. ticular tissues from the ipsilateral testis. On the basis of gross It is now well known that the action of an estrogen on appearance, tumor and “normal” tissues were separately dis- target cells is mediated by its interaction and subsequent acti- sected, and each piece was divided. One half of each sample was vation of ERs, which are members of the steroid and thyroid snap frozen in liquid nitrogen and stored at Ϫ80°C until ready hormone receptor superfamily (19). After activation, these re- for studies of receptor transcript expression. The other half was ceptors, in association with a myriad of coactivators and repres- frozen in OCT embedding compound (Triangle Biomedical sors, act as nuclear transcription factors for targeted genes (20). Sciences, Durham, NC) from which frozen sections were cut to From a teleological standpoint, the action of a hormone in a confirm the diagnoses. Histological study revealed that 3 were given tissue is substantiated by the identification of its specific pure seminomas, 1 was classified as a mixed germ cell tumor receptor in that tissue. with embryonal cell components (embryonal ϩ seminoma), and Two ER subtypes, -␣ (the classical ER) and -␤ (a recently 1 was a mixed tumor with endodermal cell components (yolk discovered subtype), have been identified and shown to exhibit sac ϩ seminoma). functional differences as well as tissue-/cell type-specific ex- Immunohistochemistry. Primary antibodies used in- pression (21). Recent investigations have demonstrated differ- cluded GC17 polyclonal rabbit anti-ER-␤ (Biogenex, San ential expression of ER-␣ and ER-␤, and their variants in the Ramon, CA), mouse monoclonal anti-ER-␣ (Nova Castra, human normal testis (22). Whereas ER-␣ has been reported to Newcastle-upon-Tyne, United Kingdom), rabbit polyclonal be strongly expressed in the efferent ductules, it has not been anti-AR (Upstate Biotechnologies, Lake Placid, NY), and rabbit localized in testicular germ line or somatic cells of human testis polyclonal anti-PR (Santa Cruz Biotechnology, Santa Cruz, (22). In contrast, ER-␤ immunostaining has been observed CA). The GC17 polyclonal rabbit anti-ER␤ primary antibody recently in spermatogonia, spermatocytes, and spermatids, as has been fully characterized in our laboratory (29). The antibody well as somatic cells, leading to the conclusion that the ER-␤ recognizes the first 22 amino acid of the F-domain of the subtype is the principal mediator of estrogen action in promot- wild-type ER-␤, exhibits high specific affinity for the wild-type ing germ cell survival and development (17, 22–25). receptor, and shows no cross-reactivity with ER-␣. Whereas the strong expression of ER-␤ in normal male The slide-mounted sections were baked for1hat60°C, germ cells is now established, analogous information is not then deparaffinized with two xylene washes. They were rehy- available for testicular germ cell tumors. Earlier studies that drated through a series of graded alcohol washes and finally used radioligand-binding methodologies for ER detection rinsed in water. Heat-induced antigen retrieval was performed in lacked the ability to discern which receptor subtype was ex- 0.01 M citrate buffer (pH 6.0). After the sections returned to pressed in testicular cancers (26–28). Similarly, immunohisto- room temperature, they were washed with water. Endogenous

Table 1 Sequences of primers for RT-PCR analysis of targeted transcripts, localization of the primers in the coding sequences, and sizes of the expected PCR products Targeted mRNA Primer sequence Location Size (bp) ER-␤ ER-␤-1:5ЈTGA AAA GGA AGG TTA GTG GGA ACC3Ј nt.230–253 528 ER-␤-2:5ЈTGG TCA GGG ACA TCA TCA TGG3Ј nt.737–757 ER-␣ ER-␣-1:5ЈTAC TGC ATC AGA TCC AAG GG3Ј nt.41–60 650 ER-␣-2:5ЈATC AAT GGT GCA CTG GTT GG3Ј nt.671–690 PR PR-1:5ЈGAT TCA GAA GCC AGC CAG AG3Ј nt.1817–1836 533 PR-2:5ЈTGC CTC TCG CCT AGT TGA TT3Ј nt.2330–2349

peroxidase was blocked with 3% hydrogen peroxide in metha- lished literature or designed using the Primer3 Output program.5 nol for 30 min, and the slides were then rinsed and set in PBS. Primer sequences for ER-␣, ER-␤, PR, AR, and GAPDH are To reduce nonspecific antibody binding, the slides were incu- given in Table 1. PCR conditions were optimized for quantifi- bated for 10 min in Power Block (Biogenex) at room tempera- cation of relative message contents under nonsaturating condi- ture, then rinsed copiously in water and again set in PBS. Before tions. Preliminary experiments were conducted to ensure line- application of the primary antibody, the sections were incubated arity for all of the semiquantitative procedures. Hot start PCR in 10% normal goat serum for 15 min. The 10% goat serum was using AmpliTaq Gold DNA polymerase (Perkin Elmer) was shaken off, and the sections were incubated overnight at 4°C used in all of the amplification reactions. The enzyme was with the primary antibody. Slide washing was performed by first activated by preheating the reaction mixtures at 95°C for 6 min before PCR. This protocol was chosen to minimize nonspecific briefly rinsing in PBS, then gently agitating for 20 s in Optimax product amplification. The routine PCR program was 30 cycles detergent (Biogenex), and by finally thoroughly rinsing with of 1 min at 94°C, 1 min at 60°C (annealing temperature), and 1 PBS for 10 min on a rotator platform. The sections were min at 72°C with the following modifications: (a) amplification incubated with the secondary antibody for 30 min at 37°C, and for ER-␤ cDNA used an annealing temperature of 58°C; (b) the wash sequence was repeated. The sections were incubated amplifications of ER-␣ cDNA and AR cDNA were carried out with avidin and biotinylated horseradish peroxidase from the at an annealing temperature of 55°C; (c) cycle number for ER-␣ Vectastain Elite ABC kit (Vector Laboratories, Burlingame, cDNA amplification was 35; and (d) cycle-number for GAPDH CA)for1hatroom temperature, and the wash sequence was was 28. Amplification of the correct sequence was verified by again performed. Staining was visualized using Sigmafast 3, direct DNA sequencing of each PCR product from at least two 3Ј-diaminobenzidine (Sigma, St. Louis, MO). The slides were different samples. then rinsed with water, counterstained with 10% hematoxylin, Statistical Analyses. One-way ANOVA was used to and progressively dehydrated through graded alcohols and analyze whether there was a significant difference among the xylene. various group means. A multiple range test using the Tukey-B Human prostate cancer sections found previously to procedure was used to compare the individual group means. express ER-␤ and AR (29) were used as positive controls for Significance was set at P Ͻ 0.05. the GC17 antibody. ER-␣- and PR-positive human breast cancer specimens were used as positive controls for the ER-␣ RESULTS and PR antibodies, respectively. All of the sections were Immunohistochemistry; ER-␤ independently reviewed by two pathologists (J. Z. and I. L.), Normal Testes. In the 5 specimens of normal testicular and immunostaining was qualitatively graded as weak, mod- tissues and in the majority of morphologically normal tissues erate, or strong. adjacent to cancers, intratubular germ cells demonstrated high RNA Isolation and RT-PCR. Tumor and normal tissues expression of ER-␤ (Fig. 1A). Specifically, strong expression of from the 5 paired testicular specimens prospectively collected the receptor was evident in spermatagonia, primary and second- were homogenized and extracted for total tissue mRNA using ary spermatocytes, but less so in elongate spermatids. ER-␤ the RNA Stat-60 reagent (Tel-Test Inc., Friendwood, TX) ac- immunostaining was in general absent in Sertoli cells (Fig. 1A) cording to protocols provided by the manufacturer. The quality Seminoma. In 14 cases, cancer cells were arranged in of each total RNA sample was checked and controlled by the diffuse sheets separated by thin delicate septae visible at higher following steps: (a) measurement of absorbance; (b) running of magnification. In 3 cases, foci of intratubular seminoma were a denaturing RNA gel capable of detecting possible RNA deg- also present. The overall intensity of ER-␤ expression was radation, as judged by the integrity and intensity of the 18S and reduced or absent in the majority of tumor cells when compared the 28S rRNA signals; and (c) conducting a semiquantitative with expression in germ cells of the normal testicular tissues in RT-PCR for the 18S rRNA at low cycle numbers. One ␮gof total cellular RNA was reverse-transcribed using the GeneAmp RNA PCR kit (Perkin-Elmer, CT), and 2 ␮l of the resulting cDNA was used in each PCR. 5 Internet address: http://www.genome.wi.mit.edu/cgi-bin/primer/ Intron-spanning primers were either obtained from pub- primer3.cgi.

Fig. 1 A, a normal human seminiferous tubule immunostained for ER-␤. Note the strong nuclear staining for the receptor in spermatogonia (double arrow) and in primary, secondary spermatocytes, and round spermatids. Staining is absent in Sertoli cells (single arrow), ϫ450. B, typical immunostaining of ER-␤ in a seminoma. Left, in this low-power field lightly stained cells are seen interspersed with negatively stained tumor cells, ϫ250. In a minority of seminomas receptor staining of tumor cells was more abundant and stronger than seen in this example. Right, the lightly stained tumor cells are better visualized in this representative higher magnification of the left panel. Staining appears to be localized in dispersed nuclear euchromatin of tumor cells. Compare this staining pattern and intensity with the strong even immunostained nuclei of normal testicular germ cells, ϫ425. C, ER-␤ immunostaining in a endodermal sinus component (yolk sac) of a mixed tumor. Note the strong nuclear staining in the tumor cells, ϫ250. D, representative section of an embryonal carcinoma immunostained for ER-␤. Staining for the receptor is completely absent in embryonal cells forming a solid pattern. The same staining was found when embryonal components were present in mixed tumors, ϫ425. E, ER-␤ immunostaining in a mature teratoma. Note the strong nuclear uniform staining for the receptor in cells that have formed cartilage as well as in surrounding undifferentiated stromal cells. Strong receptor immunostaining was also observed in epithelial cells (not shown in this field), ϫ275. F, a normal human epididymal tubule for ER-␣. Strong nuclear staining is evident in most cells lining the tubule. In contrast, staining for ER-␣ was absent in all cells of the normal testes and in all testicular germ cell cancers, ϫ300. Counterstained with hematoxylin.

Fig. 2 A, a section of a seminiferous tubule adjacent to a seminoma. B, a section of a seminoma. Both sections have been immunostained for PR. A, note the absence of staining in Sertoli cells (arrows) and the presence of strong receptor expression in spermatogonia, ϫ325. B, a small focus of PR-positive tumor cells within a seminoma. In all seminomas positive staining for PR was restricted to small aggregates or single tumor cells. The vast majority of neoplastic cells were negative for PR staining, ϫ450. Counterstained with hematoxylin.

12 of the 17 seminoma specimens (Fig. 1B). In the remaining 5 cells of the epididymis (Fig. 1F). In 3 specimens there was specimens only 25–50% of the tumor cells expressed the ER-␤, sufficient tissue available to also study the localization of AR whereas in most cells receptor staining was absent. When nu- and PR. AR was highly expressed in 2 of 3 specimens where it clear pleomorphism was present in a seminoma (anaplastic was localized to the nuclei of primary and secondary spermato- cells), a marked total absence of ER-␤ staining was strikingly cytes, and mature spermatocytes, but not in the nuclei of sper- evident in these atypical tumor cells. No differences in staining matagonia. In addition, weak AR immunostaining was also intensities were evident when solid tumors were compared with present in Sertoli and Leydig cell nuclei. PR immunostaining intratubular cancers. Variable receptor staining of lymphocytic was only detected in the nuclei of spermatagonia (Fig. 2A). infiltrates was found in the majority of cases. Seminoma. ER-␣ was not expressed in any seminoma Endodermal Sinus (Yolk Sac Tumors). Five mixed cells. AR was also not visualized in any of the tumor cells. In 4 germ cell tumors had an endodermal sinus component. In all 5 cases, a few immunopostive PR cells were found scattered of the cases, either a reticular or a glandular pattern predomi- among fields of negatively stained tumor cells (Fig. 2B). nated in the endodermal sinus component. Irrespective of the Endodermal Sinus Tumors. Staining for ER-␣ or PR pattern in this component, there was strong expression of ER-␤ was not present in neoplastic cells. There was trace AR expres- in neoplastic cells (Fig. 1C). sion in cancer cells of 3 of 5 cases. There was no expression of Embryonal Carcinoma. Both solid and glandular pat- PR seen in any of the tumors or in cells of adjacent normal terns were found in all 3 of the cases. In solid areas immuno- testicular tissues. staining for ER-␤ was either absent or very weakly positive in Embryonal Carcinoma. Irrespective of the pattern areas where tumor cells are arranged in a solid pattern (Fig. 1D), formed by embryonal cells, ER-␣, AR, or PR staining was not but was moderate in regions where glands were formed. present. Mature Teratoma. Strong ER-␤ immunostaining was Mature Teratoma. No ER-␣, AR, or PR immuno- present in all 4 of the specimens, and was of equal intensity in staining was demonstrated in neoplastic cells. stromal and epithelial components (Fig. 1E). Mixed Germ Cell Tumors. ER␣ and AR were also Mixed Germ Cell Tumors. Ten mixed cell tumors had absent in these malignant germ cells. There was trace expression both seminoma and embryonal cell components. In these cases, of PR in seminoma cells in 4 of 9 cases, but staining was absent comparison with normal testicular germ cells showed that im- in all of the embryonal cell components. munostaining for ER-␤ was weak in seminoma cells and greatly reduced or completely absent in embryonal cells. Receptor Transcript Levels Measured by Semiquantitative RT-PCR Immunohistochemistry; ER-␣, AR, and PR Semiquantitative RT-PCR was used to assess the relative Normal Testes. There was no expression of ER-␣ in any abundance (target mRNA signal over signal of GAPDH) of somatic or germ cell component in these specimens. However, ER-␤, PR, AR, and ER-␣ in total mRNA extracts of 3 pure strong nuclear staining for both ER subtypes was present in the seminomas and 2 mixed germ cell tumors (with significant nuclei of efferent tubular epithelial cells and in tubular principal seminoma cell component), and compared with relative receptor

Fig. 3 Semiquantitative RT-PCR analyses of (A) ER-␤,(B)PRin3 seminomas (cancer 1–3) and 2 mixed germ cell tumors (cancer 4 and 5), and in their corresponding ipsilateral normal testes. Total RNA was isolated and subjected to semiquantitative RT-PCR. The values of relative mRNA abundance were expressed as ratios of signal intensity of PCR products of the target cDNA over the corresponding GAPDH signal intensity. Histograms are mean values of three independent RT-PCR analyses of transcripts of ER-␤, PR, or AR in this set of normal and cancerous samples. Bars, ϮSDs. denotes significant differences ء between levels in normal and cancerous tissues at P Ͻ 0.05.

transcript abundance in the ipsilateral normal testes in the same cryptorchidism and hypospadias to declining sperm counts (30). patients. ER-␤ mRNA was found to be expressed in all 5 of the The localization of ER-␤ in normal testicular germ cells repre- normal testes, but expression was notably reduced (approxi- sents a potential mechanism through which these environmental mately Ͼ50%) in the 5 tumor specimens (Fig. 3A). Similarly, estrogens affect male reproductive function at the testicular PR mRNA was present in normal testicular specimens but was level. barely detectable in all 5 of the tumors (Fig. 3B). In contrast, Although in utero or adult exposure to estrogens has been relative AR mRNA abundance was comparable between normal implicated in male germ cell carcinogenesis, little is known of and cancerous testicular specimens (Fig. 3C). its mechanism of action. Expression of ER-␤ in spermatogonia, the purported precursor of germ cell cancers, is of particular DISCUSSION interest, because these cells lie outside of the blood testis barrier Despite past epidemiological findings that linked estrogens maintained by the tight junctions of Sertoli cells (31). Without to the development of testicular germ cell tumors (see “Intro- the protection afforded by this barrier, spermatogonia, unlike duction”) research in this area was hampered, in part, by the other germ cells within the seminiferous tubules, are subjected inability to localize ER in the testis or testicular neoplasms. This to influences of the systemic milieu. Significant levels of ER-␤ led many investigators, in the past, to conclude that ERs do not expression in spermatogonia may offer protection against cir- play a role in either normal germ cell function or malignant culating carcinogens, because this receptor subtype has been transformation (26, 27). showed to regulate specific detoxification enzymes (32). Con- In 1996, a second ER subtype, termed ER-␤, was discov- versely, ER-␤ may mediate estrogen-induced carcinogenenicity, ered and found to be highly expressed in the testes and in tissues because exposure of rat testis to estrogen was found to result in of the male reproductive tract (21). In the current study, we have oxidative DNA damage, as evident from accumulation of 7,8- confirmed recent findings that ER-␤, but not ER-␣, is expressed dihydro-8-oxo-2Ј-dexoyguanosine in testicular cell nuclei, via in normal human testis (22). Moreover, we find, in agreement an ER-mediated process (18). with others (24), that ER-␤ is strongly expressed in spermato- In the current study, we consistently detected decreased gonia, and primary and secondary spermatocytes. Because these expression of both ER-␤ RNA and protein in seminomas, mixed cell types do not express ER-␣, any direct effects of estrogen on cell tumors, and in embryonal carcinomas when compared with them are likely mediated by ER-␤. Intense attention has also normal testes. Down-regulation of ER-␤ action is additionally been focused on the widespread exposure of gonadal cells to suggested by decreased PR mRNA and protein expression seen environmental “endocrine disruptors.” Such disruptors, includ- in the same tumor-types, because the PR gene is known to be ing xenoestrogens and estrogen mimics, have been identified as regulated by estrogen (33). In contrast, the immunohistochem- possible culprits in male reproductive disorders, ranging from ical expression of ER-␤ remains strong in endodermal sinus

tumors and in mature teratomas. Conceptually, the differences 7. Strohsnitter, W. C., Noller, K. L., Hoover, R. N., Robboy, S. J., in ER-␤ expression among seminomas, embryonal carcinomas, Palmer, J. R., Titus-Ernstoff, L., Kaufman, R. H., Adam, E., Herbst, and other NSGCT tumors may be explained by the Tetrahedron A. L., and Hatch, E. E. Cancer risk in men exposed in utero to diethylstilbestrol. J. Natl. Cancer Inst., 93: 545–551, 2001. model of histogenesis (3). This model proposes that all germ cell 8. Dieckmann, K. P., Endsin, G., and Pichlmeier, U. How valid is the tumors arise from a common precursor called the IGCNU. prenatal estrogen excess hypothesis of testicular germ cell cancer? A IGCNU cells are considered to be transformed cells that arise case control study on hormone-related factors. Eur. Urol., 40: 677–683, directly from germinal cells. After transformation, the IGCNU 2001. cells give rise to seminomas, which are believed to be the 9. Olsson, H. L., Bladstrom, A., and Alm, P. Gynecomastia and risk for precursors of all other germ cell cancers. Implied in this scheme malignant tum. BMC. Cancer (Phila.), 2: 26, 2002. is that cytological and immunohistochemical differences noted 10. Hem, E., Attramadal, A., and Tveter, K. J. Synchronous bilateral between seminomas and NSGTs represents the acquisition primary germ cell tumors in patient receiving estrogen therapy. Urology, and/or deletion of traits seen as this neoplasm evolves into other 31: 70–71, 1988. distinct tumor types. Viewed in this context, the down-regula- 11. Fleming, L. E., Bean, J. A., Rudolph, M., and Hamilton, K. Cancer ␤ incidence in a cohort of licensed pesticide applicators in Florida. tion of ER- seen in seminomas, the complete lost of receptor J. Occup. Environ. Med., 41: 279–288, 1999. expression in embryonal cell tumors, and the persistent strong 12. Ohlson, C. G., and Hardell, L. Testicular cancer and occupational expression in endodermal sinus tumors and teratomas may exposures with a focus on xenoestrogens in polyvinyl chloride plastics. therefore reflect divergent pathways of differentiation/dediffer- Chemosphere, 40: 1277–1282, 2000. entiation as tumor cells progress in their development. 13. Bosland, M. C. Hormonal factors in carcinogenesis of the prostate Whereas the functional consequences of differential ER-␤ and testis in humans and in animal models. Prog. Clin. Biol. Res., 394: expression among different germ cell tumors are not currently 309–352, 1996. known, data from animal models and human cell culture studies 14. Carroll, P. R., Whitmore, W. F., Jr., Herr, H. W., Morse, M. J., suggest that ER-␤ may control and limit cell proliferation during Sogani, P. C., Bajorunas, D., Fair, W. R., and Chaganti, R. S. Endocrine and exocrine profiles of men with testicular tumors before orchiectomy. breast, prostate, and colon cancer progression (34–36). Taken J. Urol., 137: 420–423, 1987. together, these considerations and our findings have led us to 15. Uysal, Z., and Bakkaloglu, M. Serum estradiol as a tumour marker hypothesize that exposure to estrogens or estrogen-mimics, in for non-seminomatous germinal cell tumours (NSGCT) of the testis. Int. some as of yet undefined manner, diminishes the ER-␤-medi- Urol. Nephrol., 19: 415–418, 1987. ated growth restraint in spermatagonia, which favors unsched- 16. Kotula-Balak, M., Wojciechowska, J., and Bilinska, B. The site of uled cell proliferation. The affected spermatogonia or their aromatization in the mouse cryptorchid testis. Folia Biologica-Krakow, 49: 279–283, 2001. descendants may then be able to escape normal cell cycle 17. Carreau, S., Bourguiba, S., Lambard, S., Galeraud-Denis, I., Genis- regulation and be at a higher risk of undergoing malignant sel, C., and Levallet, J. Reproductive system: aromatase and estrogens. transformation. Furthermore, the diminution of ER-␤ expression Mol. Cell Endocrinol., 193: 137, 2002. in seminomas, the purported precursors of NSGCTs, its near 18. Wellejus, A., and Loft, S. Receptor-mediated ethinylestradiol- total silencing in embryonal tumors, and the re-expression of the induced oxidative DNA damage in rat testicular cells. FASEB J., 16: receptor in endodermal sinus tumors and teratomas may reflect 195–201, 2002. the progressive evolution of these germ cell cancers into various 19. Evans, R. M. The steroid and thyroid hormone receptor superfam- types. ily. Science (Wash. DC), 240: 889–895, 1988. In conclusion, our findings provide a link by which estro- 20. Katzenellenbogen, B. S., Montano, M. M., Ediger, T. R., Sun, J., Ekena, K., Lazennec, G., Martini, P. G., McInerney, E. M., Delage- gens and estrogen-like compounds may influence the genesis Mourroux, R., Weis, K., and Katzenellenbogen, J. A. Estrogen recep- and progression of testicular germ cell cancers through their tors: selective ligands, partners, and distinctive pharmacology. Recent interaction with ER-␤. Prog. Horm. Res., 55: 163–193, 2000. 21. Kuiper, G. G., Enmark, E., Pelto-Huikko, M., Nilsson, S., and Gustafsson, J. A. Cloning of a novel receptor expressed in rat prostate REFERENCES and ovary. Proc. Natl. Acad. Sci. USA, 93: 5925–5930, 1996. 1. Lee, F., Hamid, R., Arya, M., and Patel, H. R. Testicular cancer: 22. Saunders, P. T., Millar, M. R., Macpherson, S., Irvine, D. S., current update and controversies. Hosp. Med., 63: 615–620, 2002. Groome, N. P., Evans, L. R., Sharpe, R. M., and Scobie, G. A. ER␤1 and 2. McGlynn, K. A. Environmental and host factors in testicular germ the ER␤2 splice variant (ER␤cx/␤2) are expressed in distinct cell cell tumors. Cancer Investig., 19: 842–853, 2001. populations in the adult human testis. J. Clin. Endocrinol. Metab., 87: 3. Ulbright, T. M., Amin, M. B., and Young, R. H. Chap 2. Germ cell 2706–2715, 2002. tumors: histogenetic considerations and intratubular germ cell neoplasia. 23. Pentikainen, V., Erkkila, K., Suomalainen, L., Parvinen, M., and In: J. Rosai and L. H. Sobin. Armed Forces Institute of Pathology. Atlas Dunkel, L. Estradiol acts as a germ cell survival factor in the human of Tumor Pathology. Third Series, Fascicle 3: Tumors of the Testis, testis in vitro. J. Clin. Endocrinol. Metab., 85: 2057–2067, 2000. Adnexa, Spermatic Cord and Scrotum, pp. 41–180, Armed Forces 24. Makinen, S., Makela, S., Weihua, Z., Warner, M., Rosenlund, B., Institute of Pathology, Washington, DC, 1999. Salmi, S., Hovatta, O., and Gustafsson, J. K. Localization of oestrogen 4. Henderson, B. E., Benton, B., Jing, J., Yu, M. C., and Pike, M. C. receptors ␣ and ␤ in human testis. Mol. Hum. Reprod., 7: 497–503, Risk factors for cancer of the testis in young men. Int. J. Cancer, 23: 2001. 598–602, 1979. 25. Oliveira, C. A., Zhou, Q., Carnes, K., Nie, R., Kuehl, D. E., 5. Depue, R. H., Pike, M. C., and Henderson, B. E. Estrogen exposure Jackson, G. L., Franca, L. R., Nakai, M., and Hess, R. A. ER function during gestation and risk of testicular cancer. J. Natl. Cancer Inst., 71: in the adult male rats. Endocrinology, 143: 2399–2409, 2002. 1151–1155, 1983. 26. Murphy, J. B., Emmott, R. C., Hicks, L. L., and Walsh, P. C. 6. Weir, H. K., Marrett, L. D., Kreiger, N., Darlington, G. A., and Estrogen receptors in the human prostate, seminal vesicle, epididymis, Sugar, L. Pre-natal and peri-natal exposures and risk of testicular germ- testis, and genital skin: a marker for estrogen-responsive tissues? J. Clin. cell cancer. Int. J. Cancer, 87: 438–443, 2000. Endocrinol. Metab., 50: 938–948, 1980.

27. Bojar, H., Weissbach, L., Petzinna, D., Maar, K., and Staib, W. 32. Pettersson, K., and Gustafsson, J. A. Role of estrogen receptor beta Steroid receptor status of human testicular tumors. Urol. Int., 40: 160– in estrogen action. Annu. Rev. Physiol., 63: 165–192, 2001. 163, 1985. 33. Gronemeyer, H., Meyer, M. E., Bocquel, M. T., Kastner, P., Tur- 28. Kommoss, F., Franklin, W. A., and Talerman, A. Estrogen and cotte, B., and Chambon, P. Progestin receptors: isoforms and antihor- progesterone receptors in endodermal sinus (yolk sac) tumor. Evaluation mone action. J. Steroid Biochem. Mol. Biol., 40: 271–278, 1991. of immunocytochemical and biochemical methods. J. Reprod. Med., 34: 34. Roger, P., Sahla, M. E., Makela, S., Gustafsson, J. A., Baldet, P., 943–945, 1989. and Rochefort, H. Decreased expression of estrogen receptor beta pro- 29. Leav, I., Lau, K. M., Adams, J. Y., McNeal, J. E., Taplin, M. E., tein in proliferative preinvasive mammary tumors. Cancer Res., 61: Wang, J., Singh, H., and Ho, S. M. Comparative studies of the estrogen 2537–2541, 2001. receptors ␤ and ␣ and the androgen receptor in normal human prostate 35. Pasquali, D., Rossi, V., Esposito, D., Abbondanza, C., Puca, G. A., glands, dysplasia, and in primary and metastatic carcinoma. Am. J. Bellastella, A., and Sinisi, A. A. Loss of estrogen receptor beta expres- Pathol., 159: 79–92, 2001. sion in malignant human prostate cells in primary cultures and in 30. Sharpe, R. M., and Skakkebaek, N. E. Are oestrogens involved in prostate cancer tissues. J. Clin. Endocrinol. Metab., 86: 2051–2055, falling sperm counts and disorders of the male reproductive tract? 2001. Lancet, 341: 1392–1395, 1993. 36. Weyant, M. J., Carothers, A. M., Mahmoud, N. N., Bradlow, H. L., 31. Bart, J., Groen, H. J., van der Graaf, W. T., Hollema, H., Hendrikse, Remotti, H., Bilinski, R. T., and Bertagnolli, M. M. Reciprocal expres- N. H., Vaalburg, W., Sleijfer, D. T., and de Vries, E. G. An oncological sion of ER␣ and ER␤ is associated with estrogen- mediated modulation view on the blood-testis barrier. Lancet Oncol., 3: 357–363, 2002. of intestinal tumorigenesis. Cancer Res., 61: 2547–2551, 2001.

Vernon Pais, Irwin Leav, Kin-Mang Lau, et al.

Clin Cancer Res 2003;9:4475-4482.

Updated version Access the most recent version of this article at: http://clincancerres.aacrjournals.org/content/9/12/4475

Cited articles This article cites 32 articles, 4 of which you can access for free at: http://clincancerres.aacrjournals.org/content/9/12/4475.full#ref-list-1

Citing articles This article has been cited by 3 HighWire-hosted articles. Access the articles at: http://clincancerres.aacrjournals.org/content/9/12/4475.full#related-urls

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Subscriptions Department at [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://clincancerres.aacrjournals.org/content/9/12/4475. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.